def test_project_run(self):
p = self._create_project()
p_id = p['project_id']
upload_response_1 = rest_client.project_upload_clks(
self.url, p_id, p['update_tokens'][0], self.clk_data_1)
upload_response_2 = rest_client.project_upload_clks(
self.url, p_id, p['update_tokens'][1], self.clk_data_2)
run_response = rest_client.run_create(self.url,
p_id,
p['result_token'],
0.999,
name='clkhash rest client test')
assert 'run_id' in run_response
r_id = run_response['run_id']
with pytest.raises(ServiceError):
rest_client.run_get_status(self.url, p_id, 'invalid-run-id',
p['result_token'])
with pytest.raises(ServiceError):
rest_client.run_get_status(self.url, p_id, r_id, 'invalid-token')
status1 = rest_client.run_get_status(self.url, p_id, r_id,
p['result_token'])
assert 'state' in status1
assert 'stages' in status1
print(rest_client.format_run_status(status1))
# Check we can watch the run progress this will raise if not
# completed in 10 seconds
for status_update in rest_client.watch_run_status(
self.url, p_id, r_id, p['result_token'], 10, 0.5):
print(rest_client.format_run_status(status_update))
# Check that we can still "wait" on a completed run and get a valid
# status
status2 = rest_client.wait_for_run(self.url, p_id, r_id,
p['result_token'], 10)
assert status2['state'] == 'completed'
coord_result_raw = rest_client.run_get_result_text(
self.url, p_id, r_id, p['result_token'])
coord_result = json.loads(coord_result_raw)
assert 'mask' in coord_result
assert len(coord_result['mask']) == 1000
assert coord_result['mask'].count(1) > 10
result_a = json.loads(
rest_client.run_get_result_text(
self.url, p_id, r_id, upload_response_1['receipt_token']))
result_b = json.loads(
rest_client.run_get_result_text(
self.url, p_id, r_id, upload_response_2['receipt_token']))
assert 'permutation' in result_a
assert 'rows' in result_a
assert 1000 == result_b['rows'] == result_a['rows']
rest_client.run_delete(self.url, p_id, r_id, p['result_token'])
rest_client.project_delete(self.url, p_id, p['result_token'])
def test_project_run_before_data(self):
p = self._create_project()
p_id = p['project_id']
upload_response_1 = rest_client.project_upload_clks(self.url, p_id, p['update_tokens'][0], self.clk_data_1)
run_response = rest_client.run_create(self.url, p_id, p['result_token'], 0.999, name='clkhash rest client test')
with pytest.raises(ServiceError):
json.loads(rest_client.run_get_result_text(self.url, p_id, run_response['run_id'], upload_response_1['receipt_token']))
def upload(clk_json, project, apikey, server, output, verbose):
"""Upload CLK data to entity matching server.
Given a json file containing hashed clk data as CLK_JSON, upload to
the entity resolution service.
Use "-" to read from stdin.
"""
if verbose:
log("Uploading CLK data from {}".format(clk_json.name))
log("To Entity Matching Server: {}".format(server))
log("Project ID: {}".format(project))
log("Uploading CLK data to the server")
response = project_upload_clks(server, project, apikey, clk_json)
if verbose:
log(response)
json.dump(response, output)
def test_upload_clks(self):
schema_object = json.load(open(SAMPLE_DATA_SCHEMA_PATH, 'rt'))
p = self._create_project(schema=schema_object)
upload_response = rest_client.project_upload_clks(self.url, p['project_id'], p['update_tokens'][0], self.clk_data_1)
assert 'receipt_token' in upload_response